Challenges in forecasting infectious disease outbreaks
IDM conference, 6 November 2024
https://epiforecasts.io
Short-term forecasts can inform decision making
Anticipate healthcare demand
Short-term forecasts can inform decision making
Design interventions
Short-term forecasts can inform decision making
Plan clinical trials
Initial efforts faced multiple challenges
1-week ahead forecasts produced by SPI-M for SAGE in the UK.
Forecast hubs supported systematic collection of forecasts
Median ensemble outperformed individual models
Case forecasts were poorly calibrated a few weeks from the forecast date
Unpredictable changes in human behaviour made forecasting harder
Measurements of behaviour changes did not improve predictions
Observed behaviour as predictor: improvement of forecasts, but only once age-specific reporting is taken into account.
Variants as predictor improved forecasts during transitions
https://github.com/epiforecasts/forecast.vocs
https://github.com/jbracher/branching_process_delta
Humans were better than models at predicting cases, but not deaths
No clear pattern of which model performed best
Sherratt et al., in progress
For forecasts to be available when it most matters we need models and infrastructure ready to be deployed.
Charniga et al., Epidemics, 2024
(for more on mpox forecasting see also Will Green’s poster)
Any model of the future is a prediction and can be evaluated as such.
Slides at
https://epiforecasts.io/slides/idm_20241106.html